Can Vision-Language Models Solve the Shell Game? | Tiedong Liu, Wee Sun Lee | ResearchPod